单臂临床试验生存分析的样本量估计

Sample size estimation for survival analysis in single-arm trials

目的提出并验证单臂临床试验进行生存分析时的样本量估计方法。方法通过理论公式的推导,得到单臂临床试验进行生存分析时样本量估计的计算公式,并采用STATA软件进行蒙特卡洛随机模拟,估计实际的检验效能,从而验证公式的准确性。结果随机模拟结果表明在α=0.025和α=0.01的情况下,用本文提出的公式计算得到的样本量模拟得到的检验效能与预设的检验效能基本一致。结论在单臂临床试验中利用本文的公式计算得到的样本含量能够达到预定的检验效能。本研究的成果为抗肿瘤药物的早期临床研究的样本量估计提供了可靠的依据。

Objective To propose a method to calculate sample size for survival analysis in single-arm trials. Methods Through theoretical formula derivation, we provided sample size estimation method for survival analysis in single-arm trials. Monte-Carlo simulations using STATA software were conducted to obtain the empirical power, so as to verify the accuracy of the estimated sample size. Results Simulation results showed that, when α = 0. 025 and α = 0.01, the power of the sample size calculated by the formula is in accordance with the predetermined power. Conclusions The formula of sample size estimation we proposed could achieve the predetermined power for survival analysis in single-arm trials. Findings of this study provide reliable basis for sample size estimation of early stage antineoplastic drug clinical studies.